{"id":"https://openalex.org/W7154576679","doi":"https://doi.org/10.48550/arxiv.2604.13759","title":"The cognitive companion: a lightweight parallel monitoring architecture for detecting and recovering from reasoning degradation in LLM agents","display_name":"The cognitive companion: a lightweight parallel monitoring architecture for detecting and recovering from reasoning degradation in LLM agents","publication_year":2026,"publication_date":"2026-04-15","ids":{"openalex":"https://openalex.org/W7154576679","doi":"https://doi.org/10.48550/arxiv.2604.13759"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.13759","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.13759","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2604.13759","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5133751787","display_name":"Rafflesia Khan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Khan, Rafflesia","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5133824951","display_name":"Nafiul Islam Khan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Khan, Nafiul Islam","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12808","display_name":"Ferroelectric and Negative Capacitance Devices","score":0.15970000624656677,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T12808","display_name":"Ferroelectric and Negative Capacitance Devices","score":0.15970000624656677,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11636","display_name":"Artificial Intelligence in Healthcare and Education","score":0.1160999983549118,"subfield":{"id":"https://openalex.org/subfields/2718","display_name":"Health Informatics"},"field":{"id":"https://openalex.org/fields/27","display_name":"Medicine"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}},{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.06599999964237213,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/cognitive-architecture","display_name":"Cognitive architecture","score":0.5860000252723694},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5853000283241272},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.5688999891281128},{"id":"https://openalex.org/keywords/cognition","display_name":"Cognition","score":0.5055999755859375},{"id":"https://openalex.org/keywords/proxy","display_name":"Proxy (statistics)","score":0.5012999773025513},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.47600001096725464},{"id":"https://openalex.org/keywords/degradation","display_name":"Degradation (telecommunications)","score":0.3467999994754791},{"id":"https://openalex.org/keywords/scale","display_name":"Scale (ratio)","score":0.33809998631477356}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6884999871253967},{"id":"https://openalex.org/C20854674","wikidata":"https://www.wikidata.org/wiki/Q4386060","display_name":"Cognitive architecture","level":3,"score":0.5860000252723694},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5853000283241272},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.5688999891281128},{"id":"https://openalex.org/C169900460","wikidata":"https://www.wikidata.org/wiki/Q2200417","display_name":"Cognition","level":2,"score":0.5055999755859375},{"id":"https://openalex.org/C2780148112","wikidata":"https://www.wikidata.org/wiki/Q1432581","display_name":"Proxy (statistics)","level":2,"score":0.5012999773025513},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.47600001096725464},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.383899986743927},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.37779998779296875},{"id":"https://openalex.org/C2779679103","wikidata":"https://www.wikidata.org/wiki/Q5251805","display_name":"Degradation (telecommunications)","level":2,"score":0.3467999994754791},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.33809998631477356},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.3224000036716461},{"id":"https://openalex.org/C123657996","wikidata":"https://www.wikidata.org/wiki/Q12271","display_name":"Architecture","level":2,"score":0.32100000977516174},{"id":"https://openalex.org/C21200559","wikidata":"https://www.wikidata.org/wiki/Q7451068","display_name":"Sensitivity (control systems)","level":2,"score":0.3124000132083893},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.30169999599456787},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.2969000041484833},{"id":"https://openalex.org/C2780598303","wikidata":"https://www.wikidata.org/wiki/Q65921492","display_name":"Flexibility (engineering)","level":2,"score":0.2906000018119812},{"id":"https://openalex.org/C200601418","wikidata":"https://www.wikidata.org/wiki/Q2193887","display_name":"Reliability engineering","level":1,"score":0.28780001401901245},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.2800000011920929},{"id":"https://openalex.org/C120936955","wikidata":"https://www.wikidata.org/wiki/Q2155640","display_name":"Empirical research","level":2,"score":0.27489998936653137},{"id":"https://openalex.org/C113843644","wikidata":"https://www.wikidata.org/wiki/Q901882","display_name":"Interface (matter)","level":4,"score":0.274399995803833},{"id":"https://openalex.org/C66882249","wikidata":"https://www.wikidata.org/wiki/Q169336","display_name":"Homogeneous","level":2,"score":0.26460000872612},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.25870001316070557}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.13759","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.13759","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2604.13759","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.13759","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"language":[1],"model":[2],"(LLM)":[3],"agents":[4],"on":[5,20,66,76,92,105,131,151,162,183],"multi-step":[6],"tasks":[7,94],"suffer":[8],"reasoning":[9],"degradation,":[10],"looping,":[11],"drift,":[12],"stuck":[13],"states,":[14],"at":[15,118],"rates":[16],"up":[17],"to":[18],"30%":[19],"hard":[21,26],"tasks.":[22,165],"Current":[23],"solutions":[24],"include":[25],"step":[27],"limits":[28],"(abrupt)":[29],"or":[30,160],"LLM-as-judge":[31],"monitoring":[32,45,212],"(10-15%":[33],"overhead":[34],"per":[35],"step).":[36],"This":[37],"paper":[38,192],"introduces":[39],"the":[40,87,179,191],"Cognitive":[41],"Companion,":[42,103],"a":[43,54,61,112,132,171,197,202,220,230],"parallel":[44],"architecture":[46],"with":[47,70,97],"two":[48],"implementations:":[49],"an":[50,71],"LLM-based":[51,88],"Companion":[52,89],"and":[53,80,153,224],"novel":[55],"zero-overhead":[56],"Probe-based":[57,102],"Companion.":[58],"We":[59],"report":[60],"three-batch":[62],"feasibility":[63,198],"study":[64,199],"centered":[65],"Gemma":[67],"4":[68],"E4B,":[69],"additional":[72],"exploratory":[73],"small-model":[74,167],"analysis":[75],"Qwen":[77],"2.5":[78],"1.5B":[79],"Llama":[81],"3.2":[82],"1B.":[83],"In":[84],"our":[85],"experiments,":[86],"reduced":[90],"repetition":[91],"loop-prone":[93,152],"by":[95],"52-62%":[96],"approximately":[98],"11%":[99],"overhead.":[100],"The":[101,205],"trained":[104],"hidden":[106],"states":[107],"from":[108],"layer":[109],"28,":[110],"showed":[111],"mean":[113],"effect":[114],"size":[115],"of":[116],"+0.471":[117],"zero":[119],"measured":[120,180],"inference":[121],"overhead;":[122],"its":[123],"strongest":[124],"probe":[125],"result":[126],"achieved":[127],"cross-validated":[128],"AUROC":[129],"0.840":[130],"small":[133],"proxy-labeled":[134],"dataset.":[135],"A":[136],"key":[137],"empirical":[138],"finding":[139],"is":[140],"that":[141,210],"companion":[142,227],"benefit":[143],"appears":[144],"task-type":[145,217],"dependent:":[146],"companions":[147,175],"are":[148,158],"most":[149],"helpful":[150],"open-ended":[154],"tasks,":[155],"while":[156],"effects":[157],"neutral":[159],"negative":[161],"more":[163],"structured":[164],"Our":[166],"experiments":[168],"also":[169],"suggest":[170],"possible":[172],"scale":[173],"boundary:":[174],"did":[176],"not":[177],"improve":[178],"quality":[181],"proxy":[182],"1B-1.5B":[184],"models,":[185],"even":[186],"when":[187],"interventions":[188],"fired.":[189],"Overall,":[190],"should":[193],"be":[194,214],"read":[195],"as":[196,219,229],"rather":[200],"than":[201],"definitive":[203],"validation.":[204],"results":[206],"provide":[207],"encouraging":[208],"evidence":[209],"sub-token":[211],"may":[213],"useful,":[215],"identify":[216],"sensitivity":[218],"practical":[221],"design":[222],"constraint,":[223],"motivate":[225],"selective":[226],"activation":[228],"promising":[231],"direction":[232],"for":[233],"future":[234],"work.":[235]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-04-17T00:00:00"}